gnus-html.el (gnus-html-wash-tags): Limit end-tag matching to the current line to...

diff --git a/lisp/gnus/ChangeLog b/lisp/gnus/ChangeLog
index 85150a1adbf08ecfd6d8d2657f326a594385bd94..509ad305d544d4a9179d328cb9fb8455ee8785f4 100644 (file)
--- a/lisp/gnus/ChangeLog
+++ b/lisp/gnus/ChangeLog
@@ -1,5 +1,8 @@
 2010-09-06  Lars Magne Ingebrigtsen  <larsi@gnus.org>
 
+       * gnus-html.el (gnus-html-wash-tags): Limit end-tag matching to the
+       current line to work around bugs in the output from w3m.
+
        * gnus-async.el (gnus-async-article-callback): Always prefetch images
        for groups that want that.
 

diff --git a/lisp/gnus/gnus-html.el b/lisp/gnus/gnus-html.el
index eaa0e99436eae5ab6e3c3cfe086cdd0b64eee6de..034d65ef1c27e22bd85f9d826f46dc8b3e3dacc4 100644 (file)
--- a/lisp/gnus/gnus-html.el
+++ b/lisp/gnus/gnus-html.el
@@ -117,7 +117,7 @@ fit these criteria."
     (while (re-search-forward " *<pre_int> *</pre_int> *\n" nil t)
       (replace-match "" t t))
     (goto-char (point-min))
-    (while (re-search-forward "<a name[^>]+>" nil t)
+    (while (re-search-forward "<a name[^\n>]+>" nil t)
       (replace-match "" t t))
     (goto-char (point-min))
     (while (re-search-forward "<\\([^ />]+\\)\\([^>]*\\)>" nil t)
@@ -127,7 +127,7 @@ fit these criteria."
       (when (plusp (length parameters))
        (set-text-properties 0 (1- (length parameters)) nil parameters))
       (delete-region start (point))
-      (when (search-forward (concat "</" tag ">") nil t)
+      (when (search-forward (concat "</" tag ">") (line-end-position) t)
        (delete-region (match-beginning 0) (match-end 0)))
       (setq end (point))
       (cond
@@ -224,7 +224,7 @@ fit these criteria."
     (goto-char (point-min))
     ;; The output from -halfdump isn't totally regular, so strip
     ;; off any </pre_int>s that were left over.
-    (while (re-search-forward "</pre_int>" nil t)
+    (while (re-search-forward "</pre_int>\\|</internal>" nil t)
       (replace-match "" t t))
     (when images
       (gnus-html-schedule-image-fetching (current-buffer) (nreverse images)))